A detailed methodology for the long-term in vitro culture and analysis of three-dimensional, self-structuring bone models generated from cell lines or primary osteoblastic cell populations

Choice of cell source

This method (originally described in Ref. 19) relies on osteoblast-mediated contraction of a biocompatible scaffold (fibrin hydrogel) around a source of calcium and phosphate ions in the form of two cement anchors consisting of a combination of brushite (CaHPO₄.2H₂O) and B-TCP (Ca₃(PO₄)₂). Contraction provides a 3D environment within which the osteoblasts, over time, replace the original fibrin matrix with organotypic collagen and facilitate its calcium phosphate mineralisation.

The original method^19,20 utilised primary rat periosteal cells. The figures included here were generated using the human osteoblastic cell line, hFOB 1.19 (RRID: CVCL_3708) – selected based on their homogenous human origin, as well their stable karyotype,²¹ but the SSBM method can be readily adapted to accommodate osteoblasts derived from a number of different species, from primary osteoblasts isolated from vertebrate bone samples or from cell lines. We have used primary human osteoblasts and the human osteoblast cell line, hFOB 1.19,²² with similar success (data not shown). We note that primary cells have high levels of batch-batch variation in cell type populations and metabolic rates, as has been previously observed²³ and that there are intrinsic differences in the proliferation rates of cell lines therefore, seeding densities and culture media composition should be adapted and validated on a case-by-case basis.

The following preparation, set up and maintenance methods have been separated into 5 distinct aspects, summarised in Figure 1.

1. Anchor preparation

Note – Sections 1.1-1.3 are prepared under non-sterile conditions

1.1 Anchor mould preparation

Required:

Steps:

Figure 2. Preparing and using a mould for creating beta-tricalcium phosphate (β-TCP) cement anchors.

(A) A 3D sketch of a positive frame for the anchor mould, containing twenty equally spaced pegs, is drawn up with the dimensions provided above, using computer-aided design software, and a reusable frame is 3D printed with polylactic acid. (B) Dental impression material is pressed into the positive frame, ensuring the whole frame is filled, after which the mould is peeled out once completely set. (C) The recommended set up when preparing the β-TCP cement anchors. An ice tray is prepared with a spare anchor mould containing a single insect pin in each well placed in a top corner (black dotted box) and an empty mould in the adjacent bottom corner; as well as the orthophosphoric acid (3.5 M) in the vacant top corner. The β-TCP is weighed out and kept to one side of the weighing boat and placed in the vacant bottom corner of the ice tray (yellow dotted box) at a slight forward angle. The orthophosphoric acid is added to the same weighing boat as the β-TCP but kept separate prior to mixing. (D) When mixed, the resultant cement created, using the weight and volumes outlined [Section 1.3] is sufficient to fill 2-3 rows of the empty anchor mould, and pins from the spare anchor mould (C, black dotted box) are inserted immediately, prior to hardening (i). After repeating cement preparation, the cement is left to completely set within the anchor mould for approximately 3 days (ii).

1.2 Orthophosphoric acid (3.5 M) preparation

Required:

Steps:

1.3 Calcium phosphate anchor preparation

Required:

Note - The recommended workspace set-up is shown [Figure 2.C.i] to ensure required equipment is within reach during anchor preparation.

Steps:

2. Plate preparation

Note – Plates may be prepared under non-sterile conditions, as they are sterilised just prior to hydrogel preparation.

Required:

Steps:

Figure 3. Plate and hydrogel preparation, cell seeding and construct maintenance.

(A) The elastomer base, followed by the curing agent, are weighed at a 10:1 ratio into the same weighing boat, and then thoroughly mixed with the end of an empty syringe. The mixture is then transferred to an empty 6-well culture plate using this same syringe and left to set for 7 days. (B) Once the elastomer has set, 2 β-TCP cement anchors are inserted into each well of the 6-well plate. Firstly, 2 marks are made 15 mm apart on the underside of the plate for each well (i). The pins within the β-TCP anchors are bent at a slight angle, towards the long edge of the trapezoidal base using tweezers (ii), then the pins are inserted at an angle into the elastomer (white arrows), using the measured points as guide, so that the base of the anchors sit on the top surface of the elastomer (white dotted line) (iii). (C) SSBM set up involves preparing the hydrogel by addition of the thrombin across each well of the plate, and ensuring the whole surface is covered by gently mixing and tapping the sides of the plate. Addition of fibrinogen straight into the centre of each well begins polymerisation and creation of the hydrogel. Cells are seeded directly on top of the polymerised hydrogel and cultured under normal culture conditions. (D) 3 days post-cell seeding, a scalpel is run around the walls of each well to detach any fibrin from the edge of the culture well. (i). When pipetting from the well, such as during media renewal, the plate is tilted forward to pool the liquid at the bottom to ensure that the fibrin matrix is not disturbed (ii). Created with Biorender.com.

3. Reagent preparation for the fibrin hydrogel

3.1 Phosphate buffered saline (PBS) solution

Required:

Steps:

3.2 Fibrinogen solution (20 mg/mL)

Required:

Steps:

3.3 Aprotinin (10 mg/mL)

Required:

Steps:

3.4 Aminohexanoic acid (200 mM)

Required:

Steps:

3.5 Thrombin (200 U/mL)

Required:

Steps:

4. Preparing the fibrin hydrogel and construct set up

Required:

Steps:

5. Maintenance of constructs

Required:

Steps:

3-days post cell seeding – fibrin detachment

This step aids fibrin contraction by detaching any fibrin that may have adhered to the well walls

Expected results post-cell seeding

The first significant observation is contraction of the fibrin matrix around the anchors [Figure 4.A.] which, depending on the cell source used, can occur in as little as 24 hours or up to 7 days post cell seeding. Cell-matrix interactions between the cell-bound integrins and Arg-Gly-Asp (RGD)-containing moieties, such as vitronectin, found within the extracellular matrix (ECM),²⁴ are essential to cell behaviour, phenotype and survival. 2D culture platforms typically adhere cells through weak electrostatic forces on one side of the cell, which encourages an unnatural flattened conformation. By seeding cells in a 3D, naturally occurring matrix, the strong affinity between these integrins and ECM components on all surfaces of the cells results in complete enrobing of the cells in the matrix, which is observed as matrix contraction.²⁵ Furthermore, with the presence of the two anchors in the well, the matrix contraction is controlled and forced into a cylindrical-like structure, whereas a disorganised spheroid will form in their absence.²⁵ Because of the cylindrical matrix, the cells themselves are organised into a parallel and striated conformation [Figure 4B], which provides an in-built strain that further encourages cell differentiation and bone formation.

Figure 4. Matrix and cell organisation over time.

(A) Once cells are seeded onto the fibrin hydrogel, contraction of the matrix around the two calcium phosphate anchors (green arrows) occurs. Upper and lower panels are the same images, with construct outline (dotted red line) excluded in the lower panel for clarity. Over time, the cells progressively form an organised and coordinated construct (as indicated by the narrowing of the red dotted rings). The resulting structure has a defined border-due to fibrin accumulation – which becomes stiffer and increasingly opaque following introduction of osteogenic conditions at 4 weeks. (B) Immunofluorescence staining of the cell nuclei (blue) and actin (red) shows that the cells present within the fibrin matrix align longitudinally between the two anchors, resulting in a striated configuration. This Z-stack maximum intensity projection image is of a construct cultured for 4-weeks. 20x magnification, scale bar is 100 μm. Stains: Hoechst (cat no.: 62249; supplier: Fisher Scientific) for nuclei (blue), phalloidin-488 (cat no.: A12379; supplier: Fisher Scientific) for actin (red). (C) Initially, mineral deposits are observed as dark specks (white dotted rings) of varying size within the matrix of the SSBM constructs (i). Following introduction of osteogenic media at 4 weeks, these become denser and more apparent (ii). Calcium phosphate anchor within red dotted ring. Images taken with light microscope; 4x magnification. Scale bars are 200 μm.

Over time, the fibrin matrix contracts into a more defined structure, which then becomes increasingly opaque upon culturing under osteogenic conditions. This opacity is due to the occurrence of osteoid deposition by the cells and its subsequent mineralisation on the matrix [Figure 4.C].

Analytical efficiency

With the correct planning, many analytical approaches can be carried out using a single construct, which demonstrates the analytical efficiency of the SSBM construct method; for example, Figure 5 shows the journey of a single construct through five separate analytical techniques. First the construct is stained intact using non-toxic ReadyProbes™ (cat no.: R37609; supplier: Invitrogen, Hertford, UK) for live-dead imaging. Subsequently, it is subjected to X-ray fluorescence (XRF) and then scanning electron microscopy (SEM). Meanwhile, the media collected from these constructs can be analysed by enzyme-linked immunosorbent assays (ELISA) at multiple timepoints throughout the construct culture period. Constructs stained with the ReadyProbes™ are also in an acceptable condition to be used for other analytical techniques, such as qPCR and immunofluorescence (IF) staining.

Figure 5. Achieving efficient analysis of SSBM cultures.

(i) For analysis of exposed SSBM construct media, 3-4 days prior to the chosen analytical timepoint, intact SSBM constructs are isolated from their culture plates by cutting a strip of elastomer surrounding it, (ii) then incubated it in a fresh culture plate. At the timepoint, media can be collected and used for analysis, such as enzyme-linked immunosorbent assay (ELISA). The remaining construct can also be used for further analysis, such as live-dead imaging with ReadyProbes™ (cat no.: R37609; supplier: Invitrogen). Greater analytical efficiency can be achieved by using live-dead imaged constructs for other analytical approaches. To date, methods that have been optimised, and are under development, for SSBM construct analysis include quantitative polymerase chain reaction (qPCR), x-ray fluorescence (XRF), scanning electron microscopy (SEM) and immunofluorescent (IF) staining. Created with Biorender.com.

SSBM isolation

The protocol above details the measures used to ensure cells attach to and remain within the fibrin matrix, and prevent cell adherence and proliferation occurring outside of the fibrin matrix – i.e. using hydrophobic Sylgard 184 silicone elastomer and regularly treating the wells with trypsin-EDTA. Although hydrophobic surfaces won’t completely prevent cell attachment (as previously demonstrated by the murine osteoblastic cell line, MC3T3), they can reduce cellular proliferation rate.²⁶ Furthermore, trypsin-EDTA is a calcium chelator,²⁷ which could affect the development of calcium-phosphate mineral, therefore it is used sparingly throughout the culture period. Given this, two separate cell populations exist within the culture dish - one outside and the other inside the SSBM construct. This is problematic for methods that involve analysing exposed media, (e.g. ELISA) as the results will reflect the activity of both cell populations, rather than just those within the SSBM construct. To overcome this problem, prior to media analysis, SSBM constructs must be isolated from their original culture plates and transferred into fresh plates. This can be achieved whilst retaining their shape integrity, and viability, by dissecting the elastomer as close as possible to the SSBM construct, with the anchors still attached, and placing it into a new 6-well dish [Figure 5]. Following culture in the appropriate media (i.e. complete or osteogenic media) for 3-4 days, the media is collected and used for analysis immediately, or stored for analysis at a later date (refer to manufacturer’s guidance for any specific preparation steps or storage requirements for cell supernatant, prior to beginning analysis or storing) – long-term storage at -80 °C is adequate for most analysis.

Biological analysis

Live-dead imaging

There are many approaches that can be used to determine cell viability based on their enzymatic and metabolic activity (e.g. MTT tetrazolium reduction, resazurin reduction and protease activity assays), or their membrane integrity (e.g. trypan blue exclusion and DNA binding dyes). However, these methods are often destructive, irreversible or cause long-term cytotoxicity. ReadyProbes™ penetrate the SSBM construct matrix and enable simple, as well as fast viability testing, without causing cell damage or cytotoxicity. Following manufacturer guidance, constructs still attached to their anchors and a strip of elastomer (as indicated above [Figure 5]) are resuspended and incubated for 30 min (33.5°C or 37 °C, 5% CO₂) in 3.5 mL media with 2 drops/mL of NucBlue^® Live and NucGreen^® Dead reagents added directly, then imaged using a fluorescence microscope with standard DAPI (excitation/emission: 360/460 nm) and FITC/GFP filters (excitation/emission: 504/523 nm).

RNA extraction for qPCR

Quantitative polymerase chain reaction (qPCR) is a widely used technique that enables gene expression analysis in cells and tissue samples. During osteoblast differentiation, from progenitor cells through to osteocytogenesis, there are notable and patterned changes in the expression of specific genes, which have been reviewed extensively.^28–30 Data from qPCR can be used as an indicator of osteoblast cell maturity within the SSBM construct, as well as to provide insights into how cells are responding to particular stimuli (e.g. bone formation modulators) and culture environment etc.

Analysis by qPCR (or RNA sequencing) relies on extraction of RNA of satisfactory quantity and quality, which is greatly influenced by the first steps in the extraction method; disruption and homogenisation of the cells and/or tissue lysates. This method has been developed and optimised for extracting good quality RNA from the SSBM constructs [Figure 6A], using the QIAGEN Mini RNEasy^® Kit (cat no.: 74104; supplier: QIAGEN, Manchester, UK).

Figure 6. Optimised approaches for structural analysis of SSBM constructs.

(A) SSBM constructs are disrupted and homogenised prior to extraction of RNA for qPCR analysis, by following a process optimised for these constructs. Disruption involves vortexing samples, without the anchors, with glass beads, then by repeated aspiration using a syringe and needles. Disrupted samples are homogenised using QIAShredders (cat no.: 79656; supplier: QIAGEN), with the resultant sample ready for RNA extraction. (B) The elemental composition of the mineral present on SSBM constructs can be analysed using x-ray fluorescence (XRF). Whole constructs are first snap frozen in a dry ice bath using isopentane, immediately followed by dry freezing. XRF produces elemental map read-outs and energy spectra which can be used to locate and determine mineral distribution and maturity. The elemental maps shown here reveal co-localisation of calcium (red) and phosphorous (green), resulting in a yellow colouration, suggesting presence of a calcium phosphorous compound (yellow) i.e. bone-like mineral. Β-TCP anchors within white dotted outline. Scale bar is 2.0 mm. Created with Biorender.com.

First, the SSBM construct must be unhooked from the anchors, then placed it in a 2.0 mL eppendorf tube containing 10-15 3 mm diameter glass beads (cat no.: Z143928-1EA; supplier: Sigma-Aldrich) and 600 μL of buffer RLT containing 10% β-mercaptoethanol (v/v) (cat no.: M6250; supplier: Sigma-Aldrich). Each sample is vortexed (bench top vortex) for 1-2 minutes, followed by shredding by repeated aspiration with a syringe first using a 16G (cat no.: 11532445; supplier: Fisher Scientific) needle, then a 19G needle (cat no.: 15341537; supplier: Fisher scientific), the construct and the buffer RLT mixture is repetitively aspirated with the syringe until the construct is a fine powder in suspension. Care is needed as a vacuum is created in the presence of large pieces of construct. To reduce RNA loss, the same syringe is used between needles for the same sample; but to prevent cross contamination, a new syringe and set of needles are used for each individual SSBM construct. The resultant lysate is transferred to a QIAshredder spin column (cat no.: 79654; supplier: QIAGEN) and centrifuged (2 min; max speed in a microcentrifuge), followed by a repeat centrifugation of the resultant supernatant, with the shredder removed and collection tube capped. The supernatant is transferred by careful pipetting - to ensure no pelleted debris is taken up - to a new 1.5 mL Eppendorf and then used for RNA extraction following the steps outlined by the QIAGEN RNEasy^® Mini handbook, beginning with addition of 70% ethanol to the supernatant (section “Protocol: Purification of Total RNA from Animal Tissues”, pages 51-52, as of October 2019). Typical RNA yields for hFOB 1.19 are shown in Table 3, with the purity of the RNA extracted deemed acceptable when the ratio of absorbance at 260 and 280 nm is 2.0 ± 0.15.

Structural analysis

Fixation

The structure of SSBM constructs must be preserved at the elected timepoints to obtain an accurate snapshot of structural development and changes. Like bone, SSBM constructs are composite “tissues”, containing an organic and an inorganic phase, which require different methods of preservation. SSBM constructs that are to be used for structural analysis (for either the organic or inorganic phase) must first be fixed, whilst remaining attached to the anchors (as indicated above [Figure 5]). This begins with an incubation of 4.0 mL/construct 4% formaldehyde (cat no.: HT501128; supplier: Sigma-Aldrich) for 4 hours on a shaker plate at 4.0 °C; followed by three washes with 4.0 mL/construct PBS. Constructs can then be stored in PBS at 4.0 °C for up to 2 weeks. Fixation of SSBM constructs by paraformaldehyde preserves the organic phase whilst retaining the inorganic mineral phase.

Freeze-drying

For structural analysis that particularly concerns the inorganic phase of the SSBM constructs, moisture must be removed from samples. Freeze-drying (lyophilisation) is the preferred method to remove water from SSBM constructs, as it maintains sample structure (both microscopically and chemically) better than alternative approaches such as oven-drying.^31,32

SSBM constructs must first be frozen rapidly, as slow freezing can cause crystallisation of the ice, which will distort and damage the samples. This can be achieved using liquid nitrogen immersion, however, with the SSBM constructs, this can be too vigorous and can result in damage to the fragile constructs. A more successful alternative is to freeze in an isopentane bath [Figure 6.B]. Place a glass beaker, half-filled with isopentane and a small amount of dry ice, into a container of dry ice, ensuring the beaker is surrounded completely. SSBM constructs attached to their anchors and a strip of elastomer (as indicated above [Figure 5]) are placed in an Eppendorf tube and submerged in the isopentane beaker for 5 minutes. Ensure that the Eppendorf is completely submerged and add additional dry ice pellets to the isopentane if the bubbling begins to slow/stop. The submersion time can be increased if the construct is not completely frozen after 5 minutes. Once frozen, constructs are stored at -80 °C, or immediately freeze-dried (Lablyo mini, York, UK) at -50 °C for at least 4 hours. Freeze-dried constructs can be stored in a sealed container at room temperature, away from moisture, for up to 6 months.

X-ray fluorescence

X-ray fluorescence (XRF) is an inexpensive, non-destructive and rapid analytical approach to investigate the elemental composition of samples.^33–36 In this case, the location and extent of calcium and phosphorous within the SSBM constructs can be assessed. Constructs are placed in an XRF (M4 Tornado, Bruker Nano Gmbh, Germany) under a 20 mBar vacuum. After adjusting the focus on the construct, the region of interest (ROI) is outlined and then scanned with the X-ray (50 kV, 400 uA) at 1000 spots/10 mm (spot size = 25 μm), 5 ms/pixel exposure time.

By overlaying the calcium and phosphorous elemental maps [Figure 6.B], the resultant yellow colouration indicates co-localisation of the two elements, which strongly suggests the presence of calcium-phosphate mineral compounds. The resultant maps and spectra provide further information regarding the distribution and maturity of the mineral present through the varying degrees of colour intensity visualised on the maps, which are determined by the size of each element’s unique energy signal. Regarding the SSBM constructs, mineral distribution and maturation appears heterogeneous as found in in vivo bone tissue.³⁷

Immunofluorescent staining

Immunofluorescence (IF) staining techniques offer the capability to visualise the structural and biological characteristics of samples with high resolution and accuracy. The ability to section samples [Figure 7.A.i] and the vast range of fluorescent secondary antibodies available means a single sample can be analysed for numerous structural and biological markers of interest, thus furthering analytical efficiency. Furthermore, SSBM constructs can also be IF stained intact, enabling a gross overview of the cellular orientation, as well as the expression of markers of interest in relation to their location within the construct within the construct [Figure 4.B, Figure 7.A.ii].

Figure 7. Immunofluorescent staining of SSBM constructs is achievable once sectioned or whole.

Following adapted protocols for an established gelatin-based embedding, sectioning, and immunofluorescent staining protocol³⁷ (i) and one general protocol optimised for sectioned collagen-based matrices (ii), respectively, resultant 50 μm sections (i) and whole intact (ii) SSBM constructs are used to examine specific bone cell markers e.g., podoplanin (green), within the SSBM constructs, as well as an overview of cell presence and conformation. Left to right: Images above are z-stack maximum intensity projections of cultured SSBM constructs, using phalloidin for actin staining (red) and hoechst nuclei staining (blue). Scale bars are 50 μm.

Prior to staining, constructs must be fixed and then unhooked from their anchors – the constructs tend to retain their shape after unhooking once they have been fixed, as well as with increasing maturity (i.e. older constructs are stiffer).

IF Staining: Sectioned constructs

The method developed by Kusumbe et al., 2015³⁸ reliably allows the preparation of bone tissue for high-resolution IF images without compromising tissue structure or proteins. Due to the delicate and bone-like nature of the SSBM constructs, this method is also ideal for their IF staining.

The method begins with sample fixation steps, which have been adapted and described above for the SSBM constructs (refer to 'structural analysis' above), followed by decalcification and cryoprotection steps in preparation for embedding. Unlike traditional methods of embedding (e.g. paraffin embedding), this method uses gelatin-based embedding, with polyvinylpyrrolidone (PVP) and sucrose cryoprotection, as it has been found to prevent bone disintegration and allows for thicker sectioning. However, due to the fragility of the SSBM constructs in comparison to bone (i.e. a thinner composite tissue containing collagen and fibrin), the incubation conditions required for embedding are altered from 45 mins at 60 °C, to 90 mins at 45 °C. Once embedded and frozen, the protocol³⁸ outlines the steps for sectioning the resultant tissue blocks. In the case of the SSBM constructs, 50 μm sections are recommended [Figure 7.A.i]. Finally, the method outlines the IF staining procedure. Antibody incubation durations, dilutions, and fluorophore panels are subject to the researcher’s discretion. Figure 7.A.i shows SSBM constructs stained for podoplanin using an anti-human podoplanin rabbit monoclonal antibody (1:200) (cat no.: D9D7; supplier: Cell Signaling Technology, USA), followed by an Alexa Fluor^® 647 AffiniPure donkey anti-rabbit IgG secondary (1:200) (cat no.: 711-605-152; supplier: Jackson ImmunoResearch, Ely, UK). Along with hoechst (1:400) (cat no.: 62249; supplier: Fisher Scientific) for nuclei and phalloidin-488 (1:100) (cat no.: A12379; supplier: Fisher Scientific) for actin. Once stained, the SSBM construct sections are treated with ProLong™ Diamond Antifade Mountant (cat no.: P36961; supplier: Fisher Scientific) then coverslipped.

IF Staining: Whole constructs

A general IF staining protocol developed for collagen-based matrices, that have been paraffin-embedded and sectioned, has been developed in our laboratories and forms the basis of the method that is used for staining SSBM constructs whole and intact [Figure 7.A.ii].

SSBM constructs that have been unhooked from their anchors, are decalcified and cryoprotected as described by Kusumbe et al., 2015.³⁸ In a fresh 6-well plate, SSBM constructs are treated with blocking buffer [Table 4] for 1 hour at room temperature (RT) – ensuring the constructs are completely submerged (>3.0 mL solution per construct) – then washed twice with PBS. Primary antibody is diluted accordingly using antibody buffer [Table 4] and >3.0 mL solution per construct added to each well, then incubated for 2 hours at RT. Constructs are washed 3 times with PBS and treated with the secondary antibody, as well as hoechst (1:400) and phalloidin (1:100), diluted accordingly in antibody buffer [Table 4], at >3.0 mL solution per construct for 2 hours at RT and covered. Finally, the SSBM constructs are washed 3 times with PBS and are immediately imaged on a confocal microscope, not in solution, on a glass-bottomed (170±5 μm thickness) imaging dish (cat no.: 80137; supplier: Ibidi, Germany).

Troubleshooting